Hacker Newsnew | past | comments | ask | show | jobs | submit | spense's commentslogin

this looks awesome. a personal ai dom.

but i don’t trust you, so i’ll need your privacy policy before signing up. it 404s right now

https://overlord.app/privacy.html


one-shotting entire products is a dead end.

once we figure out the right building blocks of applications, it will simply be inefficient to write code yourself.

perhaps a controversial take, but writing code is going away (99% of it).


I'm extremely skeptical of this take. LLMs today aren't good at keeping significant context (even with large context windows) in such a way that doesn't lead to a spaghetti mess of code.

We're in a precarious spot where code is cheap to generate thanks to LLMs, but it's hard to build Good Software™ with only generated code. It's a pit of despair, of sorts.

If we can get LLMs to a place where that's no longer true, then writing code going away may be the new base case, but there's no guarantee we can extend LLMs that far.


we're in agreement. context rot is real.

i'm suggesting the solution is abstraction + composition. 1/ use ai to solve a subset of the problem 2/ abstract away its complexity to just its interface 3/ build with known interfaces.

this is how we solve problems. it should work the same for ai.


so many attempts at no-code solutions have come and gone.


agree, but no-code is complexity-limited bc they can't:

  1. write custom components on the fly (ai can)
  2. compose arbitrary components (composition isn't solved)


i've been looking at this recently and this isn't just for bots. ebpf fingerprinting is how cloudflare quickly detects ddos attacks.

https://blog.cloudflare.com/defending-the-internet-how-cloud...


What's the simplest way to implement eBPF filtering?

As in a NFTables/Fail2Ban level usability.



Thank you!


something like https://github.com/renanqts/xdpdropper or cilium's host firewall or https://github.com/boylegu/TyrShield exist or https://github.com/ebpf-security/xdp-firewall today and implement ebpf filter based firewalling.

Of these there is a sample integration for XDPDropper to fail2ban that never got merged https://github.com/fail2ban/fail2ban/pull/3555/files -- I don't think anyone else has really worked on that junction of functionality yet.

There's also wazuh which seems to package ebpf tooling up with a ton of detection and management components, but its not a simple to deploy as fail2ban.


Thank you


how we handle ai will dramatically shape our future.

if you consider many of the great post-ai civilizations in sci-fi (matrix, foundation, dune, culture, blade runner, etc.), they're all shaped by the consequences of ai:

   - matrix: ai won and enslaved humans. 
   - foundation: humans won and a totalitarian empire banned ai, leading to the inevitable fall of trantor bc nobody could understand the whole system. 
   - dune: humans won (butlerian jihad) and ai was banned by the great houses, which led to the rise of mentats. 
   - culture series: benign ai (minds) run the utopian civilization according to western values.
i'm a also fan of the hyperion cantos where ai and humans found a mutually beneficial balance of power.

which future would you prefer?


> i'm a also fan of the hyperion cantos where ai and humans found a mutually beneficial balance of power.

How much of the series did you read? The Fall of Hyperion makes it quite clear that the Core did not actually have humanity's best interests in mind.


Polity follows in the footsteps of Culture, with a few more shades of gray thrown in.


If I remember correctly, in Foundation they ended up heavily manipulated by a benign AI even if they thought they banned it.


Although at the very end that AI gave up control in favor of some kind of shared consciousness approach.


I'm all in on dark mode. All my devices are dark. I find it far easier on my eyes. And out of frustration that most sites don't have the option, I've enabled Chrome's auto dark mode[0], though this still breaks a few things (worth the cost to me).

[0] chrome://flags/#enable-force-dark


awesome that serde moves so quickly. i just ran across simdutf8 and realized the pr for simd-enabled uft8 parsing is coming up on 5 years:

https://github.com/rust-lang/rust/issues/68455



The parent is comparing the speed of Rust std improvement to the faster pace of Serde.


Yep. Nothing better than wowing other devs.


I mean no? The point is to share our craft and learn from each other's best techniques? The wowing part is irrelevant.


Wow then teach. If you don’t wow them, they won’t want to learn.


I think I misread the previous comment as a snarky jab. My bad!


Love these!

I'd add filling out boilerplate (e.g., impl trait defintions).

An extension can also increment selected numbers together or turn selected 0s into a sequence.


There's an inflection point where technology accrues too much power to a ruling class, such that no amount of unrest or revolution is able reset the social order.

I believe China has already passed this point. Their culture and individual behavior is tightly controlled (e.g., you can't use public bathrooms if your social score is too low).

AI may be the catalyst for western countries.


"(e.g., you can't use public bathrooms if your social score is too low)"

do you have a source on this? i have used dozens of public bathrooms in china and i've never seen any of them gated off for specific people for any reason or anyone checking anything before you go inside


The things people believe about the social credit score just shows they don't know how little we know about China. https://threadreaderapp.com/thread/1716857951380721972.html


And then there's the Uighur minority who can't buy knives that don't have QR codes on them [0], that is if they haven't been imprisoned or forced into factory labour hundreds of miles from home.

[0] https://www.fastcompany.com/40510238/in-xinjiang-china-some-...


It was the Internet. Not because the powerful gained more power, because all possible revolutionaries became opiated.


> It was the Internet. Not because the powerful gained more power, because all possible revolutionaries became opiated

Is the claim that the frequency of revolutions scales inversely with internet penetration? Because this is trivially testable and obviously false. (Ukraine and Tunisia off the top of my head.)


even here some jobs you can't get if your credit score is too low.. a definite social ramification. it leaks a bit - but it will be a sieve eventually.


If anything I’m surprised to see the ruling class isn’t starting to lose this battle to the masses. Ukraine’s use of disposable drones plainly shows how easy it is to create smart weapons that could be piloted autonomously (or at least in the “terminal” phase).


The history of technology, including weaponry, has been to increase the power of the individual with respect to the state.


Does it? Mass surveillance technology clearly only benefits the state who can run it. A random person or activist group doesn't have NSA like powers. Likewise advancements in tanks or nuclear weapons does not empower the individual.


pray tell how does the individual counter the state's use JDAMs?


By targeting the logistical chains that are necessary to deliver said JDAM to the point of actual use.

On the other hand, for the state to use JDAM against the individual, they need to know that said individual is there and is worth targeting, for starters.

I don't think it's accurate to say that the balance of power has shifted on the whole, but regular individuals definitely have access to way more destructive power than they ever had in history. At the same time, modern military technology, while very destructive, is also very demanding in terms of logistics. Even a single blown up railroad, fuel depot, large transformer etc can have a profound effect.

The destructiveness isn't always a positive, either. Any modern military can easily reduce even a very large city to rubble, but to what purpose? A city is only valuable in the grand scheme of things because of its infrastructure and its population that can utilize that infrastructure for some useful economic purpose.


look at the upside, we may be able to eliminate school shootings when the AI-powered turrets are installed in every classroom.


Nah, they'll just hack those remotely instead. :( :( :(


On the surface, China is a country where power controls behavior, but the real situation is that the Western world is


why not adam back? the evidence in favor of adam dwarfs other candidates.


honestly, i thought we settled on Adam Back as Satoshi:

  - adam was extremely active during the early cypherpunk days, when
  - he invented Hashcash, which bitcoin cites in its whitepaper and used as its PoW
  - he was the first to take Wei Dai's b-money proposal seriously and discusses combining it with hashcash (the 2 core elements of bitcoin)
  - he disappeared and reappeared professionally during the exact timeframe of bitcoin development
  - his writing closely matches satoshi's (british english + style/punctuation)
  - he had no public involvement with bitcoin until he joined Bitcointalk forums in 2013 and suddenly had intimate details of unknown bugs in bitcoin saying "I thought I'd fixed them".
  - he started Blockstream with the mission to ensure the survival of bitcoin (funded lighting, paid core devs, etc)
  - he recently helped nail the coffin on Craig Wright's case in court proving he was not satoshi, during which
  - he revealed the earliest known email correspondence with satoshi
watch the interview were he uncomfortably describes himself

https://www.youtube.com/watch?v=688J9UZJxKg

https://www.youtube.com/watch?v=XfcvX0P1b5g


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: